Two experiments examined children's ability to apply three different standards for evaluating their understanding. Five-, seven-, nine-, and eleven-year-old children were presented with short narrative passages within which were embedded three types of problems (nonsense words, internal inconsistencies, and prior knowledge violations), each of which could only be identified if a specific standard of evaluation were used (lexical, internal consistency, and external consistency, respectively). Since the focus of the study was on the effectiveness with which children could apply the standards, rather than on the likelihood that they would spontaneously adopt and then apply them, the subjects were explicitly instructed in advance that their task was to find the "mistakes." Moreover, the subjects were given immediate feedback after each trial and a second opportunity to find any missed problems. Although older children used all three standards more effectively than younger children, overall problem identification was considerably better than that reported in non-instructed settings. The internal consistency standard was applied least effectively, but even the youngest children were able to use it. The results illustrate the need to consider comprehension monitoring skills with respect to specific standards of evaluation, rather than as a unitary phenomenon. (Author)